draw-options
Skill1 starsChanged 3 months agoMITView on GitHub
Draws the options for a decision on a tldraw offline canvas - side-by-side option cards with what each includes, pros and cons, plus a comparison matrix when scoping a feature - saved to one task-named file in a single tidy folder. Use this skill whenever the user says "draw the options", "diagram the options", "pros and cons on a canvas", "compare these options visually", "sketch the approaches in tldraw", "draw options", or "/draw-options" - even if they don't name the skill. Not for editing an existing canvas in general, and not for stress-testing a design by interview (grill-with-docs).
In this repository38 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- api-contract-reviewSkill
- autopilotSkill
- backlog-plannerSkill
- codebase-explainerSkill
- code-reviewSkill
- context-dietSkill
- conventional-commitsSkill
- create-prSkill
- design-briefSkill
- diagnoseSkill
- dotnet-conventionsSkill
- dotnet-onion-apiSkill
- e2e-verifySkill
- githubSkill
- goal-runnerSkill
- grill-with-docsSkill
- handoffSkill
- humanizerSkill
- improve-codebase-architectureSkill
- maestro-mobile-testSkill
- migration-safetySkill
- model-inventorySkill
- nextjs-app-routerSkill
- plan-and-buildSkill
- record-browser-testSkill
- regression-huntSkill
- safe-refactorSkill
- security-reviewSkill
- ship-itSkill
- sql-reviewSkill
- task-executorSkill
- tauri-2-appSkill
- think-like-fableSkill
- upgrade-depsSkill
- worktree-sandboxSkill
- write-a-skillSkill
- write-testsSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 1 is distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 51
Draw: Write the spec as JSON in your scratch folder, then node <skill>/scripts/oc.mjs draw <id> columns spec.json, and the same for matrix …
- Distinctive: line 71
Leave the user's own Untitled or other canvases alone.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Draw Options
Brainstorming options is easier to judge when each one is drawn: what it includes, what is good, what is not. This skill draws them on the user's tldraw offline desktop app through a small helper, scripts/oc.mjs (<skill> below means the directory holding this file), and keeps the files tidy.
When to use this skill
- "draw the options" / "diagram the options for this task"
- "pros and cons on a canvas" / "compare these approaches visually"
- Scoping how a feature should behave and the choices are worth seeing side by side
Step 0 — preflight, always first
node <skill>/scripts/oc.mjs check decides everything that follows:
state | Meaning | Do |
ready | App installed and answering | Continue |
not-running | Installed, server not answering | Tell the user to open tldraw offline, rerun check once |
not-installed | App not found | Say so, give the options as a Mermaid diagram or table in chat, report Canvas: not run |
- ❌ Drawing "to a file" yourself, or reporting a canvas you did not see answer.
- ✅ A non-
readystate ends the canvas path; the chat fallback still answers the question.
Workflow
- Collect the options from the conversation: 2 to 4, each with a name,
includes,pros,cons. Items are short (120 characters or fewer, 12 or fewer per list). A con or pro you do not know is left out, never invented. Setrecommended(an option's exact name) and the matrixrecommendationonly when the user stated or agreed to a preference; otherwise leave both out and reportRecommended option: none.- ❌
"recommended": "One-click"because it looked simplest to you; the canvas then shows your pick as the user's. - ✅ No
recommendedfield until the user says which option they lean toward.
- ❌
Pick layouts by what you can observe:
Observable Layouts Always columnsThe user is scoping what a feature does, or you can name 3+ criteria shared by every option add matrixChoosing one option opens a follow-up decision you can name (put it in that option's thenlist)add tree- Name the file
<task#>-<short-desc>; with no task number,<yyyy-mm-dd>-<short-desc>. Runnode <skill>/scripts/oc.mjs open "<name>"and keep the returnedid. The file is created under the right name from the start, never renamed afterwards. - Draw. Write the spec as JSON in your scratch folder, then
node <skill>/scripts/oc.mjs draw <id> columns spec.json, and the same formatrix/tree. Layouts stack downward on one page. - Finish.
node <skill>/scripts/oc.mjs finish <id>saves, measures the file and takes a screenshot. Read the screenshot file before reporting.
columns spec (formats for matrix and tree are in references/protocol.md):
{"title": "Report export", "question": "How should users export a report?", "recommended": "Emailed",
"options": [
{"name": "One-click", "includes": ["Export button"], "pros": ["Fast to build"], "cons": ["Times out on big reports"]},
{"name": "Emailed", "includes": ["Request dialog"], "pros": ["Handles big reports"], "cons": ["Needs a job queue"]}]}- ❌ Placing shapes by hand with your own coordinates.
- ✅ The helper's builders: they size cards to the text and fail rather than overlap.
Files — keep them tidy
- Files live in
~/.claude/diagrams/(override with--dir), outside every repo, so nothing is committed. - A new file opens in the app by itself. One file per task: the same task again,
openwith the same name finds the file, opens it if it was closed, and the nextdrawadds below. Stateexists-not-openmeans the app did not open it within 15 seconds; ask the user to open that file, then stop. - Shapes and text only. No pasted images or screenshots on the canvas.
- Leave the user's own Untitled or other canvases alone. Never draw on a document you did not create or open here.
finishprintsbytes; ifsizeWarningis set, tell the user.- Cleanup:
node <skill>/scripts/oc.mjs listshows age and size. Delete only after the user names the file, thenremove <file> --yes. Never delete unprompted. - Nothing sensitive on the canvas: no real names, record IDs, credentials, or personal data; generic titles.
Failure protocol
draw returns ok: false: read message. bad-spec lists what to fix. layout lints remain means shorten the longest label or cell. Fix exactly that and retry once; a failed draw is rolled back, so a retry does not duplicate. A second failure: stop, report it, answer in chat.
Report — these slots are required
Draw options
- Canvas: run | not run (<reason>)
- File: <path> (<bytes> bytes)
- Layouts: columns[, matrix][, tree]
- Lints: <number from finish>
- Screenshot viewed: yes | no
- Recommended option: <name> | noneA value you did not observe is not run or no. File is filled only from a successful finish. Zero options worth drawing is a valid outcome: say so instead of padding.
Notes
- Verified on Windows only; elsewhere
checklooks in the usual app-data location and reportsnot-installedornot-runningif the app is absent. - The helper needs Node 20+ and no packages.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

